Inspiration
We wanted to provide a tool for farmers and general agricultural workers to get advance notice on when a crop they are growing could have a potential infection.
What it does
Identifies rice diseases using machine learning and image processing.
How we built it
Used Python and TensorFlow to code and design an algorithm that accurately identifies diseases in images of rice. We also created a website using HTML and CSS for farmers to use where they can submit photos of their rice plants and put it through our program to identify if it is diseased and, if so, what disease it has.
Challenges we ran into
The biggest challenge by far was learning how to implement machine learning. None of us had ever used machine learning before, and we had multiple team members working on a way to develop an algorithm and make it as accurate as possible from the ground up. Furthermore, we had a general lack of experience with coding; we are all freshmen enrolled in COMP110, and this is our first Hackathon! Another big constraint was time—learning how to use and implement HTML, CSS, machine learning, TensorFlow, Flask, and many more concepts was definitely a difficulty in such a time crunch.
Accomplishments that we're proud of
Learning TensorFlow and developing a working, accurate model that detects whether a specific infection is affecting a rice plant is definitely what we're most proud of—especially since rice is a wildly important food source for many third-world countries.
What we learned
How to do machine learning, code as a team, create websites and make great products.
What's next for Rice Savers (using Machine Learning)
We want to expand our program to do detection on many plants, not just rice; while rice is a staple for many third-world countries, having the same algorithm for crops such as wheat or different vegetables would also be invaluable. We would also like to bridge the gap between our program and our website, allowing them to interact. While we have all of the Flask code written to allow such a connection to be possible, none of us own a domain name and therefore could not host it on the internet.
Log in or sign up for Devpost to join the conversation.